I'm really too tired to argue against this in great detail, but overall:
I didn't talk only about academia, I said management in general as well. But the arguments largely apply similarly anyway.
Interests aren't due to genetics, they are due to society. An important factor among everything that biases people's interests is seeing examples in roles that might suit you. When positions are filled with people from a limited socio-demographic range, everyone else doesn't see people like them and is biased against going that career route - continuing the limited variety, and so on along the cycle.
While academic jobs won't make you rich, they pay pretty alright. It's not like people choosing against academia all go into jobs that make them rich. Education is expensive though, and pay is crap until you land an actual job. That's where finance-based affirmative action for grants and awards makes sense, to actual tap society's potential.
The accuracy of hiring processes is far overrated. You make it out as if the process will lead to a reliable ranking of candidates by quality, but that's not true. At best, you can classify candidates in categories such as 'excellent', 'good', 'okay', 'insufficient' - and even there, the application qualities of applicants (which aren't important for fulfilling the role but key to getting it) will play a role, as well as the biases of the hiring committee. It's well-known that people get ranked differently if the sociodemographic characteristics are not known. So rather than saying 'due to EDI, we pick #5 from our list and forego the superior quality of #1-4', the better characterization is that, among a group of comparably excellent candidates, the EDI candidate is chosen.
Also, studies show that there are indeed advantages for society from having more variety in all kinds of domains. For example, the 'male gaze' has meant that women's health has been under-researched for the longest time (or sickle cell disease, to take a very specific example); or creates inventions that don't work as well for people types that are too different from the developers; that entire population groups have been underserved in various ways; and so on.
Further, studies also show that companies with more women in leadership roles (more compared to severe underrepresentation) do better, on various fronts. I can't right now find the study I read previously, but there are plenty of data-driven articles about this online. For example:
https://hbr.org/2021/04/research-adding-women-to-the-c-suite-changes-how-companies-think and
https://www.forbes.com/councils/for...-everyone-wins-with-more-women-in-leadership/.